Fangcong Yin

Hi! My name is Fangcong Yin. I am a 4th-year PhD student in computer science at New York University advised by Professor Greg Durrett. I spent two and a half years at the University of Texas at Austin before transferring to NYU. Prior to that, I was an undergraduate student at Cornell University where I studied Information Science.

I am interested in research in large language models, specifically on the following topics:

Large language model post-training and agentic RL: multi-turn reinforcement learning, efficient fine-tuning
Long-context language modeling: long-context reasoning, long-horizon agents, agentic context management

I have also worked as a research scientist intern at Meta and Amazon.

I have been fortunate to work with Professor Marten van Schijndel and Professor Claire Cardie at Cornell University and Professor Meng Jiang at the University of Notre Dame.

selected publications

EMNLP

Randomized YaRN Improves Length Generalization for Long-Context Reasoning

Manas Mehta, Fangcong Yin, and Greg Durrett

In Findings of the Association for Computational Linguistics: EMNLP 2026, 2026

arXiv Bib

@inproceedings{mehta2026yarn,
  title = {Randomized YaRN Improves Length Generalization for Long-Context Reasoning},
  author = {Mehta, Manas and Yin, Fangcong and Durrett, Greg},
  booktitle = {Findings of the Association for Computational Linguistics: EMNLP 2026},
  year = {2026},
}

arXiv

Visually Grounded Self-Reflection for Vision-Language Models via Reinforcement Learning

Liyan Tang^*, Fangcong Yin^*, and Greg Durrett

Preprint, 2026

arXiv Bib

@article{tang2026reflection,
  title = {Visually Grounded Self-Reflection for Vision-Language Models via Reinforcement Learning},
  author = {Tang, Liyan and Yin, Fangcong and Durrett, Greg},
  journal = {Preprint},
  year = {2026},
}

EMNLP

Learning Composable Chains-of-Thought

Fangcong Yin, Zeyu Leo Liu, Liu Leqi, Xi Ye, and Greg Durrett

In Findings of the Association for Computational Linguistics: EMNLP 2026. Also presented at the Workshop on Foundations of Reasoning in Language Models (Oral), NeurIPS 2025 , 2025

arXiv Bib

@inproceedings{yin2025composable,
  title = {Learning Composable Chains-of-Thought},
  author = {Yin, Fangcong and Liu, Zeyu Leo and Leqi, Liu and Ye, Xi and Durrett, Greg},
  booktitle = {Findings of the Association for Computational Linguistics: EMNLP 2026},
  year = {2025},
}

NeurIPS

LoFiT: Localized Fine-tuning on LLM Representations

Fangcong Yin, Xi Ye, and Greg Durrett

In Proceedings of the 38th Conference on Neural Information Processing Systems (NeurIPS). Also presented at the Workshop on Foundation Model Interventions (Oral), NeurIPS 2024 , 2024

arXiv Bib

@inproceedings{yin2024lofit,
  title = {LoFiT: Localized Fine-tuning on LLM Representations},
  author = {Yin, Fangcong and Ye, Xi and Durrett, Greg},
  booktitle = {Proceedings of the 38th Conference on Neural Information Processing Systems (NeurIPS)},
  year = {2024},
}